Expert Data Modeling with Power BI

Soheil Bakhshi

• Editorial: Packt Publishing
• Año: 2021
• ISBN:
  9781800203174
• Categoría: Computer Science
• Categoría: Data Modelling & Design

Manage and work with business data effectively by learning data modeling techniques and leveraging the latest features of Power BI

Key Features

• Understand data modeling techniques to get the best out of data using Power BI
• Define the relationships between data to extract valuable insights
• Solve a wide variety of business challenges by building optimal data models

Book Description

This book is a comprehensive guide to understanding the ins and outs of data modeling and how to create data models using Power BI confidently.You'll learn how to connect data from multiple sources, understand data, define and manage relationships between data, and shape data models to gain deep and detailed insights about your organization.In this book, you'll explore how to use data modeling and navigation techniques to define relationships and create a data model before defining new metrics and performing custom calculations using modeling features. As you advance through the chapters, the book will demonstrate how to create full-fledged data models, enabling you to create efficient data models and simpler DAX code with new data modeling features. With the help of examples, you'll discover how you can solve business challenges by building optimal data models and changing your existing data models to meet evolving business requirements. Finally, you'll learn how to use some new and advanced modeling features to enhance your data models to carry out a wide variety of complex tasks.By the end of this Power BI book, you'll have gained the skills you need to structure data coming from multiple sources in different ways to create optimized data models that support reporting and data analytics.

What you will learn

• Implement virtual tables and time intelligence functionalities in DAX to build a powerful model
• Identify Dimension and Fact tables and implement them in Power Query Editor
• Deal with advanced data preparation scenarios while building Star Schema
• Explore best practices for data preparation and modeling
• Discover different hierarchies and their common pitfalls
• Understand complex data models and how to decrease the level of model complexity with different approaches
• Learn advanced data modeling techniques such as aggregations, incremental refresh, and RLS/OLS

Who this book is for

This MS Power BI book is for BI users, data analysts, and analysis developers who want to become well-versed with data modeling techniques to make the most of Power BI. You'll need a solid grasp on basic use cases and functionalities of Power BI and Star Schema functionality before you can dive in.

]]>

Section 1: Data Modeling in Power BI

In this section, we quickly introduce data modeling in Power BI from a general point of view. We assume you know what Power Query is, what DAX is, and that you know the basic concepts of the star schema. In this section, you will learn about virtual tables and time intelligence functionalities in DAX and how you can implement a powerful model with real-world scenarios.

This section comprises the following chapters:

• Chapter 1, Introduction to Data Modeling in Power BI
• Chapter 2, Data Analysis eXpressions and Data Modeling


Chapter 1: Introduction to Data Modeling in Power BI

Power BI is not just a reporting tool that someone uses to build sophisticated reports; it is a platform supplying a wide range of features from data preparation to data modeling and data visualization. It is also a very well-designed ecosystem, giving a variety of users the ability to contribute to their organization's data analysis journey in many ways, from sharing datasets, reports, and dashboards to using their mobile phones to add some comments to a report, ask questions, and circulate it back to relevant people. All of this is only possible if we take the correct steps in building our Power BI ecosystem. A very eye-catching and beautiful report is worth nothing if it shows incorrect business figures or if the report is too slow to render so the user does not really have the appetite to use it.

One of the most important aspects of building a good Power BI ecosystem is getting the data right. In real-world scenarios, you normally get data from various data sources. Getting data from the data sources and mashing it up is just the beginning. Then you need to come up with a well-designed data model that guarantees you always represent the right figures supporting the business logic so the report performs well.

In this chapter, we'll start by learning about the different Power BI layers and how data flows between the different layers to be able to fix any potential issues more efficiently. Then, we'll study one of the most important aspects of Power BI implementation, that is, data modeling. You'll learn more about data modeling limitations and availabilities under different Power BI licensing plans. Finally, we'll discuss the iterative data modeling approach and its different phases.

In this chapter, we'll cover the following main sections:

• Power BI Desktop layers
• What data modeling means in Power BI
• Power BI licensing considerations for data modeling
• The iterative data modeling approach

Understanding the Power BI layers

As stated before, Power BI is not just a reporting tool. As the focus of this book is data modeling, we would rather not explain a lot about the tool itself, but there are some concepts that should be pointed out. When we talk about data modeling in Power BI, we are indeed referring to Power BI Desktop as our development tool. You can think of Power BI Desktop like Visual Studio when developing an SQL Server Analysis Services (SSAS) Tabular model. Power BI Desktop is a free tool offering from Microsoft that can be downloaded from https://powerbi.microsoft.com/en-us/downloads/. So, in this book, we're referring to Power BI Desktop when we say Power BI unless stated otherwise.

The following illustration shows a very simple process we normally go through while building a report in Power BI Desktop:

Figure 1.1 – Building a new report process in Power BI

To go through the preceding processes, we use different conceptual layers of Power BI. You can see those layers in Power BI Desktop as follows:

Figure 1.2 – Power BI layers

Download the Microsoft Contoso Sales sample for Power BI Desktop from https://www.microsoft.com/en-us/download/confirmation.aspx?id=46801.

Let's discuss each point in detail:

• The Power Query (data preparation) layer
• The data model layer
• The data visualization layer

The data preparation layer (Power Query)

In this layer, you get data from various data sources, transform and cleanse that data, and make it available for other layers. This is the very first layer that touches your data, so it is a very important part of your data journey in Power BI. In the Power Query layer, you decide which queries load data into your data model and which ones will take care of data transformation and data cleansing without loading the data into the data model:

Figure 1.3 – Power Query

The data model layer

This layer has two views, the Data view and the Model view. In the Data view, you can see the data, and in the Model view, you can see the data models.

The Data view

After we are done with our data preparation in the Power Query layer, we load the data into the data model layer. Using the Data view, we can see the underlying data in our data model layer after it has been transformed in the data preparation layer. Depending on the connection mode, this view may or may not be accessible. While we can see the output of the data preparation, in this view we also take some other actions, such as creating analytical objects such as calculated tables, calculated columns, and measures, or copying data from tables.

Note

All objects we create in DAX are a part of our data model.

The following screenshot shows the Data view in Power BI Desktop when the storage mode of the table is set to Import:

Figure 1.4 – Data view; storage mode: Import

The Data view tab does not show the underlying data if the table only shows the data when the storage mode is set to Import. If the storage mode is set to DirectQuery, the data will not be shown in the Data view:

Figure 1.5 – Data view; storage mode: DirectQuery

The Model view

As its names implies, the Model view is where we stitch all the pieces together. Not only can we visually see h...